384 Mr. E. B. H. Wade. On a new Method of 



inquired what is the best relation of the distances between the 

 observer and galvanometer on the one hand, and the signal and gal- 

 vanometer on the other. Usually they are made equal, bnt a simple 

 calculation shows that for a given angular displacement of the 

 mirror the reading may be increased by bringing the observer near 

 to the galvanometer and removing the signal to as great a distance 

 as possible. Using this disposition of the galvanometer, very minute 

 deflections of the mirror were plainly visible. 



§ 5. The employment of the Apparatus to measure the difference of 

 the Boiling Points of Pure Water and Solution. 



The mode of employing the apparatus varies somewhat, according 

 as it is required to work with weak or strong solutions, and at high 

 or low pressures. The procedure with the weak solutions at 760 mm.* 

 is as follows : — Steam from the generator is caused to circulate 

 through the empty Gr-tubes and drum, and a position is found on the 

 bridge, such that on making contact the galvanometer is undeflected. 

 A reading of the bridge scale is taken and entered as "hypsometer 

 null point." Water is now added to each Gr-tube, and caused to boil 

 by the passage of steam. A similar null point is found and entered 

 as " water null point." If the burettes (Plate 2) are properly 

 adjusted the two null points are identical. In order now to measure 

 the change in the boiling point due to the presence of a small quantity 

 of salt, some water is removed from one G-tube, and replaced by a 

 few c.c. of stock salt solution. The bridge is again balanced. It was 

 found desirable that this balance should be obtained at the fifteenth 

 minute of ebullition, and therefore at this moment the bridge 

 must be correctly set, the levels duly adjusted, and the galvanometer 

 observed to be undeflected. This is a work of considerable difficulty, 

 but it may be accurately performed if all one's movements are regu- 

 lated by the indications of a clock, so that each takes place at its 

 appointed time. The observer ascertains, for instance, in a number 

 of preliminary experiments, what gain takes place in the reading 

 of the burettes during fifteen minntes, and in an actual experiment 

 the levels are previously so adjusted that they shall be as nearly as 

 possible correct at the end of such a period. The fifteen minutes 

 may now be allotted to observations at the bridge and galvanometer, 

 and then one may be sure of a close balance at the last minute. The 

 levels of the burettes are finally recorded, and a sample of solution is 

 withdrawn from the G-tubes for analysis. As a result of the pre- 

 cautions described, the burette reading was almost invariably found 

 to be within 1^ mm. of the desired height. In any case a correction 

 (L) can always be made at the rate of 0"001° per 1-J mm. inequality 

 * ' Eoy. Soc. Proc.,' vol. 61, pp. 285—287. 



